How to Have Fun With Elections; A Primer for Screwing Democrats

I am the guy that messed with American elections to help *, and here's what I had to consider and deal with to come up with my method of election fraud:

1. The method MUST provide a NET BENEFIT to * in terms of votes, at least in the targeted Electoral states, and is EXTREMELY LIKELY to provide a net benefit nationwide to insure popular vote victory.

2. Expectations are all formed by the immediately prior presidential election, which the 2004 election is shaping up as a replay of. I MUST stay within spitting distance of Bush2000's percentage performance, or have plausible cover for exceeding it.

3. If my method involves directly distorting/changing the actual reported or recorded election results as opposed to the suppression of voters prior to actually voting, the method MUST use actual votes as an input or else pollbooks showing the number of voters will be way too far away from actual counts of ballots. (i.e. we can't have 1000 votes in 500 voter precinct, but we can have 503 voters in a 500 vote district since officials will blithely conclude that 3 voters "forgot" to sign)

4. If computerized, the fraud method would BE EXTREMELY LIKELY TO use a randomizer of some sort, in order to avoid detection by well known mathematical tests for randomness.

5. My PREFERRED METHOD FOR ELECTION FRAUD WOULD BE VERY LIKELY to mix it up a little, providing some benefits for the opponent Kerry so long as the net benefit to the favored candidate *. The outliers for Kerry will give, if it ever becomes an issue, something for Rs to scream about, and the media is obligated to give at least equal time to Rs and usually more than that. As Florida2000 proved, having the better of the voting argument is no substitute for media control or court control, the first of which makes unequal things seem equal, and the second of which ensures victory.

6. To deal with exit poll problems, I would want my FRAUD TO MASQUERADE AS POLLING BIAS, (that age old bugaboo of all polls) thus totally cutting off any enterprising (and oh so PARTISAN) election activists at the knees with an army of very able and honest and completely noncomplicit statisticians and political scientists who will sincerely argue that the exit poll data is more consistent with polling bias than anything else.

SO, HERE'S HOW I SCREW THE DEMOCRATS AND JOHN KERRY:

(1) "EQUALIZE TO 2000" Using Bush 2000% performance as a baseline, (with widely published precinct figures) I would keep total votes constant in virtually all precincts I can affect, but make Bush2004% come out very close to Bush 2000%. In code, this means that where Bush2004% is less than Bush 2000%, then we use a randomizer to place Bush 2004% close to Bush 2000%, using the randomizer to keep variation in the data. Politically, this is always believable since the country has been prepped for a repeat of 2000. "ROUGHLY EQUALIZE TO 2000, NET AT 2000% plus 0.4%" or something like 0.4%.

(2) USE VOTE SWITCHING TO SPRAY NONRANDOM PRECINCTS. Based on the "repeat of 2000" scenario, the additional margin of victory is provided in large part by select precincts where touch screen DREs switch votes as widely reported across the USA. The vote switches not caught and reversed by voters are then "sprayed" randomly across the defined and non-random selection of precincts that the individual touch screen machine is assigned to serve. For example, a single corrupted touch screen DRE in an polling place containing 6 DREs switches votes right in front of the faces of the voters, and then in effect sprays those extra unearned votes across the 4-12 precincts it serves at that given polling location (each touch screen machine holding multiple ballot styles being one of the advertised "advantages" of touch screen DREs) for the simple reason that touch screen machine results are placed into various different precincts and thereafter recorded and reported to the public as "precinct" results, thus scattering the margin provided by corrupted machines over the entire polling place MIMICKING A LEGITIMATE VICTORY because the whole polling place "trends" Republican. Adding additional elegance and beauty to this method, vote switching via miscalibration of the touch screens can be caused by both accidental miscalibration as well as deliberate miscalibration, thus endlessly obfuscating issues of what causes the observed vote switches, the amount of the vote switching error is enhanced by pollworkers themselves who are trained to "blame the voter", and then when they finally do re-calibrate the touch screen voting machine the pollworker, who is totally noncomplicit, nevertheless destroys the evidence of what was producing the irregularity of vote switching in the first place. In any event, the extra switched votes are sprayed over the other precincts served by the same machine, and 99% of all the people think it's a "glitch" and that no fraudster would make their fraud visible (99% of all people plainly ignoring the "hide in plain sight" and "deniability" principles behind the need for every criminal to cover their tracks and leave false clues) VOTE SWITCHING AND SPRAYING NONRANDOM PRECINCTS.

AND/OR

(3) DEMOCRATIC VOTES ARE NOT RECORDED OR DELETED in high minority areas, establishing high undervote rates in those areas but which can be attributed strongly, as in Florida 2000, to dumb or apethetic voters who can't figure out machines and therefore probably shouldn't even be considered smart enough to vote. See New Mexico, where this happened a lot. DEMOCRATIC VOTES ARE NOT RECORDED OR DELETED

AND

(4) In recountable areas that will be close, TIME-HONORED VOTE SUPPRESSION TACTICS are used, along with(later on) recount suppression tactics if necessary, to first keep people from ever approaching the polls, then to keep them in long lines if they do approach by shorting heavily Democratic precincts from having adequate numbers of voting machines, and finally as necessary to spoil thousands of ballots by marking them with additional marks creating "overvotes" for President that don't count for anything. All of these suppression tactics in (4) are NON-RECOUNT DETECTABLE and NON-EXIT POLL DETECTABLE and otherwise invisible to exit polls except for the portion that constitutes deliberate spoilage after the voter successfully votes.

WHY THIS WORKS:

In the trench warfare where suppression is necessary or the balloting systems less penetrable or diverse such as Ohio, the discrepancies created are not detectable by a recount, nor are they detectable by exit polls because voters are supressed from seeing exit pollsters as well.

In high Dem areas and high minority areas, it's not believable for Bush to equal or exceed his 2000 performance because there are so few * supporters they could be counted up in some precincts after the election, so the preferred method is disenfranchisement through converting dem votes to undervotes, which is demonstrated in spades in New Mexico. <http://www.votersunite.org/info/NewMexico2004ElectionDataReport.pdf>

Methods are forced to adapt to diverse voting technologies, and when electronic touch screen voting is available with machines serving multiple precincts, outlier machines can be used to spray their margin (the rest are too close to call as a whole and don't provide margin), so the vote sprayers provide the margin of victory, but it looks like a regular "glitch" and in any event it will be (fallaciously) argued that every vote successfully switched the vote back to Kerry, ignoring all those that didn't see it, and this glitch aslo can provide the cover for an additional and larger outright theft of electronic votes without entry of the same by voters in the maximum amount that can be attributed to any plausible number of people who failed to spot and correct the "switch glitch" and are thus too "stupid" to have their votes properly counted anyway, like Palm Beach County Buchanan voters in 2000. The "vote switching" phenomenon is #5 on Time For Change's list recently and also has the merit of not being exit-poll-detectable. It OBVIOUSLY is not recount-detectable.

And, with regard to number (1), where election results are equalized to 2000 performance and this raises * performance, the discrepancies creatd by this tactic can plausibly be attributed to polling bias or "reluctant Bush responders". Because the exit polls will show more Kerry support than the actual election results which have been upgraded to 2000 numbers on a "rough random" basis, the discrepancy created is also consistent with rBr or polling bias, which posits that the Bush election results are real but the Bush voters shy (or Kerry voters ebullient). However, the same discrepancies are even more consistent with fraud that roughly equalizes on a somewhat randomized basis to 2000 performance.

Because statisticians are very self-conscious of polling bias issues that crop in essentially every poll, most will attribute the 2000 equalization fraud to the bias, not fully realizing that it is equally if not more valid to attribute the WPE (Within Precinct Error) to the "biases" of the election as it is to attribute it to the "biases" of the exit poll.

Finally, I suggested above that the most elegant way of all to screw the Democrats would be make it look as accidental and as Pro-Kerry as possible consistent with delivering the correct NET BENEFIT to *. This can be readily done by giving Kerry a few outliers as "gimmes". This will not destroy the net benefit to *, it gives one's own team something to complain loudly about in the press if there's ever an issue, and it destroys the "expected" correlation between redshift (Bush doing better in election result) and performance relative to 2000. In reality, one would only expect such a correlation between 2000 and 2004 if there were in fact two legitimate elections under politically stable and nonchanging environments between 2000 and 2004, and to be relatively non-correlated if there is instability politically or fraud in either election.

Remembering the above argument that a fraudster would wish to cover his tracks and obfuscate the issues by throwing a few bones to Kerry (making it look like Dem dirty tricks), take a look at this paragraph from the recent USCV report on the History of the Election Debates, focusing on Mitofsky's own findings on his own exit polls:

"An interesting finding from Mitofsky's scatter-plot is that it showed exit polled precincts in which exit polls had over-estimated the Bush vote share by amounts well beyond the margin of sampling
error. Although such over estimates of Bush vote share occurred in fewer precincts than overestimates of Kerry vote share, if such large discrepancies are being caused by vote fraud, it is not
confined to one party."

Thus, USCV correctly notes that there are these bizarre outlier precincts (and that there are more of them that favor Bush then that favor Kerry) but they fail to realize the NET BENEFIT principle, and instead suggest solely the possibility of TWO party involvement. While two party shenanigans are not only possible but likely (and would also help explain some Dem reluctance to push the issue of election fraud), USCV cites Mitofsky work without fully exhausting its possible significance, consciously describing it as "interesting". They should, and probably are, working on it more.

SO there you have it, how I'd screw the Dems and Kerry. It powerfully allies Mitofsky on my side along with a lot of statisticians, but without the slightest need for their complicity. It is truly elegant in how it gets pollworkers to first amplify the fraud with their delayed responses due to the assumption of a voter mistake, and then has the pollworker unwittingly destroy all the evidence with a recalibration. It destroys the correlations many intuitively expect between redshift and 2000 performance, but still provides NET BENEFIT to *. It allows me to focus my real calculations for the winning margin on the vote switching sprayers (a relatively finite number of nonrandom machines) knowing that the rest of the precincts will zero out to equal 2000 performance anyway, if this is even necessary in a race that was generally mimicking 2000 anyway. And, millions of Republican operatives who have no idea that I even exist will be working hard on their own suppression schemes in Ohio, which are also non-recount detectable and non-exit poll detectable.

None of my tactics will be exit poll-detectable except some of my sprays and tabulator grabs, but if they are exit poll detectable than I thow a few precincts the other way as camouflage and bait. None of them are recount-detectable (particularly the spraying electronic voting machines, so I will be laughing my ass off as nubile feminazi colege-age activists work hard to get a "recount" that won't enlighten them much at all except in a couple states. Most wonderfully, any one of these methods is probably not "enough to change the result" so the media won't care, but if they do, then there's plenty to counterattack with in terms of contrary inferences.

One final, and personal, note. A lot of you might think that I'm scared or that I was nervous or whatever doing all of this rigging. I laugh at that idea. You haven't been listening to the fact that our Western Civilization is engaged in a perpetual war with Islam that's been quiet a couple centuries but is flaring up again. I swear on my family, my country, and every single thing I hold dear in western civilization that these BARBARIANS will not destroy our civilization as they made so clear their intent to do so on 9-11. Liberal pansies like John Kerry will directly lead to the ruin of our civilization, which can not be tolerated.

So, I can tell you that the very most satisfying day of my life, and the happiest day of my life, was being the guy that threw the election for George Bush. I am the captain of my soul, and I have no doubt that history will vindicate my actions in the election of 2004. But in case there's any legal troubles in the mean time, it was all a glitch, and in any event you have no evidence, and voters are human, they make mistakes, you know???

This is like sitting down to the most anticipated meal of the season and knowing your eyes are bigger than your capacity to ingest. I read through this and it's the product of a special mind set. I'm impressed that you can immerse yourself in this, come up with the deviant deeds, and maintain your sanity.

This crime, the theft of Election 2004, is one of the most profound crimes in world history. In 2000, just as we reached our apogee and there was no counter balance, we had our country stolen by the Neocons. The 2004 election was to be the grand rescue carried out by the people. They did what they were supposed to do: registering in record numbers, giving us equal standing in funds, working their asses off for K-E and then this, the theft of all time.

I'll sleep on this and be back for more. You don't mind if I pick do you?

Great work.

A favorite saying:
"If you can keep your head,
while those around you are losing theirs,
Perhaps you don't fully understand the problem."

GO AHEAD AND MAKE MY DAY: REQUEST THE SOURCE CODE! the screen calibrations are never stored!!

There's nothing there!!!!!!!!!!!!!!!!

Rolling on the floor wickedly laughing my ass off,

I remain,

Yours truly,

LAND SHARK

the machines and their controllers after they gain access to both.

on edit,
and by the way, a lot of the smart number people are catching on to the fact that while Mitofsky sampled a random sampling of precincts nationwide, he did not sample a random sampling of the fraudster's precincts nationwide.

And if a highly funded or well connected party had even partial intelligence about what precincts would be exit polled, that would in and of itself help a lot with masking the fraud, if one then refrained in some of those precincts. In this light, both of the camps in the exit poll debate often assume that the sampling of fraudulent precincts was random (though sometimes it is consciously referred to as the possibility or probability of nonrandom fraud)

This may also undermine the exit poll's usefulness as evidence (OTOH, Febble) but you've got to go with the evidence you got. IF there is a substantial chance that democracy is under attack, the militia may need to be formed. We've got to risk some false fire alarms in order to reach all the fires. You can't sit back until you have strong and irrefutable evidence of the fire.

But I guess we all join in the need for a lot more investigation and disclosure

If any work is to be done on the machines, we have to call a technician from the electoral board and log the action that was taken. The machine may not record recalibration, but there would be a paper trail.

This may apply to just my county.

it depends on the machine too, what kind of Machines did your county use?

BTW the so-called "audit logs" of the machines I studied will simply indicate that a recalibration was done but not why or anything else about the recalibration. These audit logs are stripped of the kind of information one would want in a real audit log, i.e. it is apparently not recorded in the first place.

I am a computer programmer by profession. I took a vacation day or two to work the polls before I was laid off.
I have been the top officer in charge at my precinct and would have been the one to be trained in it if anyone was. No election officer in my precinct has ever been trained to do any maintenance other than to reboot the machine if it hangs.

This is not saying that it couldn't happen somewhere else, I only know it doesn't happen here.

If there was a problem, I would have to write specific notes as to the nature of the problem and what the resolution was. These notes would indicate the machine by serial number and be separate from any internal log in the machine. I would also have to call the headquarters and report the problem, so there would also be a phone log of the problem. If we have poll watchers from the political parties, we would get one from each party to witness what is happening. Otherwise, each precinct gets the workers to declare a party and one worker from each party serves as witness.

The fact that I have to do this should be incentive to others to work the polls and make sure that the procedures are followed as written. If I didn't care about a fair election, I could let the problem go unresolved and unreported. Make sure there is someone on the inside to see that the procedures are followed. Let that person be you.

It delays the corrective response further, by the amount of time it takes the troubleshooters to arrive. This is true regardless of the cause or nature of the problem.

We either get (1) more problems and/or misvoting or, if the machine is temporarily out of service (2) longer lines and more chaos at the polls. Sure to bring a smile to anyone who distrusts democracy.

Recalibration problems are Beautiful! Hmmm. We just need to find some way to concentrate them on Democratic polling locations. Hmmmmmmmm......

You hit the jackpot on that one.

Two precincts occupy the same polling place. One is more Dem than the other. The Repub precinct gets six machines for 2000 voters, the Dem precinct gets seven machines for 3000 voters. Small lines for Repub precinct, Long lines for Dem precinct.

Poll watchers did not notice. Poll worker complained to Elec Board. Will it make any difference? Who knows.

If you want the SOURCE CODE and all in all it doesn't matter, because the screen calibrations are never stored. Are you suggesting the only way to secure elections are PBHC?

I pooh pooh disclosure of source code in the sense that if one is defending a vehicle parked in the middle of a giant parking lot, having the source code be open source is like having a traffic barrier protecting the car from being hit, but only from the East North Easterly direction. (You're even worse off with no protection at all, it's necessary protection but nowhere near sufficient to do the job)

Elections are not TECHNOLOGY (paper or plastic defining their properties). Elections are SYSTEMS of constitutional checks and balances set in administrative realities.

So, even HCPB (while a better head start) can be readily compromised (but not so dramatically, usually)

My suggestion is to argue for certain standards and values (i.e. checks and balances) in elections systems of which HCPB happens to meet those standards, provided it also contains joint/multiple supervision by the public and opposed parties of approximately equal strength, good chain of custody, indelible markings of voter intent verified in its final form by the voter herself, etc. (HCPB under certain circumstances meets these values)

Note that these values will be almost universally agreed to. In contrast, HCPB can "feel" like a "throwback". It's not a throwback because what HCPB really is, properly configured, is a highly advanced and secure system of checks and balances for constitutionally democratic elections. THAT'S WHAT I'M FOR. At least when my evil twin isn't talking about how I rigged the 2004 election to stimulate thought.

"Note that these values will be almost universally agreed to. In contrast, HCPB can "feel" like a "throwback". It's not a throwback because what HCPB really is, properly configured, is a highly advanced and secure system of checks and balances for constitutionally democratic elections".


When your evil twin shows up he wants them vote counting and tabulating machine in between us(voter) and the vote totals, because he can steal an election, have the exit polls way out of WHACK in 22 states. Put his guy in office before the people(VOTERS) know what hit them, and/or can figure out what to do about it. He doesn't even have to sweep up all them unsightly ballots.

He is a very efficient thief ,his machine creates the perfect election crime scene not only do they commit the crime they hide all the evidence.


PBHC keeps the evidence front and center.

There were those handful of states where Kerry's count beat the exit poll. Perfect excuse for excusing the exit polls.

Thanks for bringing it all up.

They're just for you.

I accept the premise that it would be reasonable for a Bush fraud maestro (BFM) to engineer fraud in both directions if necessary to disguise the fraud. I'm not at all convinced that it happened, but there is nothing inherently wrong with the idea. (If the "Best Geo" estimates hold water, this would imply that Bush threw a net 300,000 votes to Kerry in Texas, but hey, he could certainly afford to.)

Some of the other exit poll estimates of net vote theft -- net favoring Bush -- would be 300K-600K votes in Florida, 400K-500K votes in Ohio, 850K-1 million votes in New York, and 350K-1.35 million votes in California (the Best Geo and WPE values are quite different there). Florida is the only one of these states where the official vote diverged markedly from (some) pre-election polls -- and the polls in the week before the election ranged from Kerry + 5 to Bush + 8, so who knows? If the BFM was also stealing votes on behalf of Kerry in some places, then the challenge becomes even larger (but not insuperable).

Honestly, I don't think that Bush stole one million net votes in New York, and I haven't seen anyone try to make the case. Land Shark proposes that "the additional margin of victory is provided in large part by select precincts where touch screen DREs switch votes as widely reported across the USA." Well, no. Certainly not in New York. Not in Ohio or Florida, either, as far as I can tell (remember, Kathy Dopp set out to look for touch screen fraud in Florida).

So I think the exit polls are pretty much useless. If the BFM has sufficient resources to engineer all these results at the precinct level, why not bribe some interviewers to stuff exit polls in Ohio, just to distract us from the fact that the election was actually stolen in (say) Virginia and New Mexico?

And I think the touch screen argument is problematic regardless, but worth exploring further.

the touch screen stuff is pure fact, including the spray, there's all the data you'd want. What remains is the size or magnitude of the touch screen shift, and whether it is exit-poll detectable with or without possible additional masking mechanisms (not discussed). What also remains is the accidental or intentional or combination nature of the causation of the touch screen vote switching, in terms of what's available on this thread.

Belief, even in scientists, is more powerful than fact. Some people will never believe certain things, no matter the facts (this may not be you). A family relative thinks Bush had to have won really big in NC because NC is the biggest giver to a certain religious charity, ignoring John Edwards and the data in Mark Crispin Miller's book showing discrepancies between early and absentee, on one hand, and election day on the other hand. (in other words, the family member's belief-based argument totally ignores factual data to the point of not even bothering to dismiss it)

Perhaps we can work to a point where the data is just as consistent with the nonfraud hypothesis as the fraud hypothesis that leaves little evidence, so that could lead to reform ideas..... This would require nothing more than Mitofsky's already stated statement "I don't know if there's fraud or not". But shouldn't THAT be a disturbing statement? i.e. We transferred power in the world's sole superpower and largest democracy and we don't know if it was legitimate, or not.

If my criminal imagination were more finely honed, I would probably not spend much time on this issue to begin with!

With respect to touch screens, I can only say that obviously in this context size matters, and in fact is crucially important. So, we will try to sort that out.

OMG! I didn't realize Mark's book had come out -- what planet have I been on? (Well, there hasn't been much buzz here.) Let me see what he has to say about North Carolina. (I would imagine that Bush won North Carolina reasonably big because he appears to have led in all 24 pre-election polls I have access to, by an average 8.6 points in the last five. But I'm not committed to him having won by double-digits as in the official returns.)

"...'I don't know if there's fraud or not'. But shouldn't THAT be a disturbing statement?"

Asked and answered (and answered, and answered...). My answer is yes.

in absentee (if I recall correctly), but 12% on election day. Issue is trying to account for the extra 6% and there's observed vote switching on election day. It's around page 180 in the book but don't hold me to an exact number (middle somewhere)

although I have put it down for a while. There is a lot to work with in NC because it is a crazy quilt of voting methods -- I'll have to see whether Plotner (and/or someone else) has done a lot of it already.

North Carolina is one of those states where early voting is growing apace. It increased from about 470K votes in 2000 to about one million votes in 2004, and its availability probably helps to account for North Carolina's above-average turnout increase, given that it was hardly a swing state.

My figures combine election day and early/absentee ballots, but they are still pretty suggestive (over 70% of votes still are cast on election day). Bush increased his margin of two-party vote by about 0.4 points over 2000 on op-scan, dropped 0.7 points on punchcards, dropped 1.2 points on non-touchscreen e-vote, and dropped 1.4 points on touchscreen e-vote. (Based on much smaller numbers of votes, he improved by 2.6 points on lever machines and by 5.2 points on hand-counted paper ballots.)

Now, if the early ballots had been representative of the entire electorate, Bush's overall margin would have dropped by about 7 points statewide. So, if we had some reason to insist that the early ballots were representative of the entire electorate, we might infer that Bush stole votes on every kind of equipment. I can't rule that out. But do we have any evidence for it?

Mark Crispin Miller's book (I think you were going to buy it?) says that the early voting percentages matched the exit polls for NC on election day, but the results on election day did not. pages 182-185 or so in "Fooled Again".

-- and it's important to have a "because" -- nationwide the discrepancies between the vote count and the exit polls don't correlate (or even negatively correlate) with the discrepancies between the vote count and pre-election polls. It gives me pretty much zero confidence in the reliability of the exit polls as a fraud detector.

Also because, in North Carolina, the mean Within Precinct Error was -11.3 in 2004, but also -9.8 in 2000, -6.5 in 1996, and -4.2 in 1992 -- one of the worst records of any state. That pattern could evince a steadily increasing incidence of vote count fraud (in which case someone should make the argument -- there are some very good political scientists in North Carolina, quite a few of them Democratic). Prima facie, it only means that North Carolina, like New Hampshire, was a state where based on recent experience, we would not have expected the exit polls to be reliable.

I did buy MCM's book. I didn't understand his discussion on p. 183 at all. He points to vote losses in two Republican counties as, I guess, evidence of Bush fraud (perfectly plausible, but one would want at least to say something about where the votes were lost), he points to excess votes (obviously spurious) in a third Republican county as evidence of Bush fraud, and he reports DRE problems in the "far more populous and liberal Mecklenburg County." All four of these counties apparently did have either touch-screen or punch-button e-voting, so that's why I undertook the analysis I did.

with results, that does not mean individual states don't or won't (individual state anomalies may offer canceling "noise", but on a state by state basis that noise won't nessarily exist)

If you have zero confidence in the exit polls "as a fraud detector" wouldn't you also have zero confidence in the exit polls for all purposes or other purposes? If so, wouldn't it be more accurate to say that various criteria or signs make the exit polls unreliable for several major things, of which fraud detection may be one?

Something I haven't looked into, but what's the story behind the exit poll computers going "down"? Can you rule out data damage or irregularity to the exit poll data as a result of this computer event or otherwise?

Data from individual precincts are likely to be very noisy indeed; one would hope that by pooling precincts, one would get a better idea of what is going on in a particular state. When we are discussing what we believe to be 6- or 8- or 10-point discrepancies at the state level, there is no a priori reason to expect them to wash out in state-by-state "noise." But precinct-level analysis is especially useful for variables that cut across state levels. For instance, a lot of us sort of expected the WPEs to be a lot higher in touchscreen precincts, but they aren't.

Sure, the exit polls are unreliable for several major things, including state projections, which is why the pollsters resist calling states without having some official counts to look at. When I say that I have pretty much zero confidence in the reliability of exit polls as fraud detectors, I don't mean that they convey zero information, but that they can't be relied upon in themselves.

I can't "rule out" much of anything. Heck, for all I know the Warren County Board of Elections was conducting a blood sacrifice to corrupt the Edison/Mitofsky database. But there is no obvious inconsistency between what we knew about the exit polls on election night and what we know about them now. They showed Kerry ahead then, and they show Kerry ahead now.

about withholding exit polls so as not to discourage westerners from voting. Is that what some may say a convenient excuse to have time to "adjust" or "cook" the exit poll data? Do you think it's true that we'll never see unadjusted exit poll data from here on out?

There just seems to be a whole lot of tension between the view of exit polls articulated by you and febble and their previous reputation.

Even if they are just to identify trends, the whole "moral values" trend identified and trumpeted as explanatory was a crock from the start. More people identified moral values in 2000 and 1996 than 2004, so it seems quite weakly explanatory at best.

Then there was the idea that the young people just didn't vote. Wrong.

Then hispanics were falsely claimed to have gone with shrub in a BTE (better than expected) way. Nope.

Has E-M helped in any way whatsoever? WHAT'S HE GOOD FOR?

The canonical example of an early call based on exit polls was in 1980, when apparently NBC called the election for Reagan a little after 5:00 Eastern. That was fairly rash in a way (and I personally think a bad idea), but then again, Reagan won the popular vote that year by 10 points. So it isn't so surprising that NBC got away with it.

It's important to understand, however, that the networks have never made calls based on exit polls alone (well, I dunno, someone out there might have been psychotic). Initially they are referencing against pre-election polls; if a state looks at all competitive, they wait for official returns before making a call.

The reputation of the exit polls has had a lot to do with two facts: (1) the polltakers have never relied on interview data alone, and (2) most recent elections haven't really been that close. 1992 was a relatively close one although far from a cliffhanger, and we now know that the exit poll was sort of a mess -- but they got away with it (in part because they deserved to get away with it, because they avoided wrong calls). (In 1988 at least one network botched at least one state call, but no one really cared.) 2000 was close, and obviously there was a big mistake. 2004, from Mitofsky's standpoint, there actually wasn't a big mistake, because they didn't call any of the states wrong. I'm tempted to say that the exit polls have been wrong whenever it mattered, but that's playing into the myth that the interviews are all-important, and that's not how the projections work.

There is a big difference between trying to prevent exit poll leaks at midday and trying to suppress the Call 3 estimates after the polls have closed. What I've heard is plans to hold the data close until 6 PM. No one AFAIK has stated any plan to suppress the Call 3 estimates, and I don't believe that there is one. So, no, personally I don't think it's true that we'll never see unadjusted exit poll data (assuming that you mean not adjusted to the official results).

The 2004 exit poll question that referred to "moral values" didn't match any earlier questions, so it wasn't a trend question at all. But the media sponsors paid for that question, among others, and it gave them something to talk about, so it isn't for us to say whether it was a waste of money or not. (We're entitled to our views, of course!) People are gonna be arguing about the Hispanic issue for a long time; I don't think there is any compelling evidence about what the right answer is there. But, hey, what can I say? I'm here neither to bury exit polls nor to praise them, but to contextualize them.

Nevertheless (assuming your point is true that the EXACT same question wasn't asked about moral values in the past) we still have the fact that there was no fantastic surge of moral values voters to explain the election. But it did happen to coincide nicely with R claims prior to the election that they would bring to the polls 6 million more evangelicals.

What's Edison-Mitofsky good for, again? (I am hoping that you will apply the same exacting and critical standards to any and all claims about E-M and not just to the fraud evidence claims made regarding the E-M exit polls)

Do you mean, why do the media subscribers pay for their polls? Probably the same reason that they pay for other polls, namely because it gives them something to talk and write about.

The data are IMHO useful as long as one doesn't take them literally. We are all in a tizzy because of a three-point discrepancy on Bush's vote share, but when I am studying public opinion on policy issues, the inherent error in asking people about issues they don't know much about totally swamps that three points. I've seen folks freely throw around factoids about "Bush's approval rating under 40" without even saying which pollster's data they are using, when we know darn well that it matters. And yet, it doesn't matter all that much; no pollster says that Bush's approval rating is in the 50s.

And I saw plenty of people throw around nonsense about moral values, before and after the election -- but there were also people who immediately interpreted the exit poll data in a reasonable way. Some people have better judgment than others.

If there are any claims about E-M that you would care to make or to pass along, I will do my best to be exacting, critical, and fair-minded in my assessment.

Incidentally, E-M doesn't write the questionnaires; the media sponsors decided to include "moral values" on the list of most important issues. And I think that decision was correct, because it means _something_ that a lot of people checked that box, and most of them voted for Bush.

OTOH statement: If there are any claims about E-M that you would care to make or to pass along, I will do my best to be exacting, critical, and fair-minded in my assessment.

it seems like the defense of E-M involves very substantial depreciation of E-M usefulness, which cuts both ways and ultimately undermines the integrity of E-M's data in ways that go beyond its alleged lack of ability to clearly prove election fraud.

You've answered that in part by suggesting that large differences are more meaningful.

Perhaps you've answered in substantial part, from a business/political standpoing by alluding to "media sponsors" who have a need to fill air time with analysis, pseudo or not. Also, the media's decision to include "moral values" as an important question to ask PREFIGURES to a substantial degree their later conclusion after the election that it mattered in the election results. (Unconsciously, I see them saying to themselves "See how smart us media types were for asking the moral values question in the first place?")

As you must surely know by now, I'm inviting you to find an E-M claim about their data and attack it, establishing your utter independence and rigorousness as you try to do here at DU and therefore your usefulness as a future paid consultant to E-M, perhaps ; ) If YOU suggest it, it will show your critical thinking skills are trained on E-M, and not just mine or others. (i'd prefer a new observation, not simply reminding or reciting an old one)

This is what, a hazing ritual?

Sir, I don't know what would be independent or rigorous about hunting for something I can criticize E/M about -- mind you, not something I've ever criticized them about before, oh no, something NEW -- and not necessarily anything pertinent to the discussion, as long as it is, well, critical -- just for the sake of proving that I can. I doubt you would respect me more if I accepted this invitation.

Generally speaking, as an academic, I find that E/M is not nearly as forthcoming about their research methods and data as academics (at their best) are, and this frustrates me. But of course that isn't attacking a claim they make about their data.

It seems to me that the prevailing discourse at DU has ranged from "E/M is treasonous" to "someone must have threatened Mitofsky's children" (no, I didn't make that one up, although I'm sure I'm not quoting it quite accurately). You (individually and/or collectively) are welcome to think that my independence and rigor are in question, along with Hertsgaard's intellectual honesty and Mother Jones' courage under fire. At some point I cannot afford to care what you think, sir. I am doing my best for you. If you doubt my credibility, I invite you to interview any ten qualified political scientists at random and see where they come out. It is not my problem.

As you pointed out yourself, there was much talk about evangelicals (and abortion and gay marriage and religion and "values") before the election. So it hardly required prophecy, deviousness, or even much creativity to think of adding "moral values" to the issue question. But that was obvious, yes?

Standard law school training, though I'm not deigning to condescend. It's just that a really good debater can adopt the other side and argue it well, also. So, it's not a hazing ritual so much as an invitation to harness your considerable skills in a more multifaceted and robust manner. If you could gladly accept my invitation, then you could establish both your expertise and your independence simultaneously, a formidable position.

Instead, hesitation in that department leads to the regrettable or perhaps overstated charges you describe above.

Given the fact that the sine qua non of democracy is that the public is the boss (dictators, as I've said, still like to have elections to the extent they think they can get away with them) when you have deprived the public of all meaningful oversight capability (as we have with secret vote counting) Mitofsky's data secrecy, justified or not, hurts and doesn't help democracy. The apparently untroubled nature of Mitofsky's conscience leads to inferences (I use the word "apparently" in recognition that I don't truly KNOW).

Mitofsky, whether he thinks it appropriate or not is in a position to Judge our elections, even if he thinks that the data collection method won't merit it. Perhaps he fears that responsibility. But Abdication is not the answer. It's important to know whatever is knowable, and then to expand the realm of knowledge considerably in this area. Mitofsky could, at least, be a leader in the area of data disclosure. Ballots, unlike exits, don't even arguably disclose identity because of ballot secrecy.

:thumbsup:

Thanks, as always, to you, LS!

Deciminate (this) and Conquer!

:P
I must say Land Shark, I am impressed.

It takes a special criminal mindset to delve so eloquently into the Republican strategy. The saddest part of your sickening post is that everything you said is backed up by cold hard evidence of official misconduct at the highest and lowest levels of this election.

It helps that the two swing states that effectively decided it were governed by criminally corrupt Republican regimes. Hell, Governor Taft is the only governor to ever be charged with a crime while in office in the stupefied state of Ohio. Governor Bush is....well...a Bush.

Good work Karl, I'm glad you chose a Democratic forum to leak this, especially at 12:30 AM. It helps take the heat off of that other leak problem we have. Hopefully you get enough recommended votes so more people can read this once they wake up.

I am Attorney General Alberto Gonzales and all DUers are hereby under arrest for treason. Good night and good luck. :eyes:

K+R

This is what we need.

We should go through this in depth, and add in stuff that BillBored, TfC and Yowza, and anax have come up with.

Moreover, if we can come up with some testable hypotheses that would be a bonus.

I may have issues with you over exit-poll visibility, but you are definitely on the right lines. Fraud will be hard to hide from the exit polls, but possible. We need to find the limits on the possible.

Great job.

my son's got a dentist appointment at 10, so I'm not going in to work till 11.

I'll print out your thread and cogitate while I'm waiting at the dentist's.

Mixing it up won't really help. Pro-Kerry fraud (whether perpetrated by pro-Kerry agents or pr-Bush counter-agents) will show up as blue shift AND as a reduction in Bush's "true" swing. Pro-Bush fraud will show up as red-shift AND as an increase in Bush's "true" swing.

Either way, the direction and extent of the shift will be correlated with the extent and direction of the swing.

The better approach for your fraudster is to organise it so that the algorithm only switches votes in precincts where Bush is doing badly -in other words non-random fraud.

But it will be difficult to leave fingerprints with that as well, as it should skew the swing positively (and it doesn't seem to be skewed positively). I'm working on that one though. I'd like to come up with a fool-proof scam. This is rather more fun than what I usually do.

Time to think like a criminal and what's really needed and really desirable and see how close that gets us to reported fact.

Like the transcontinental railroad, being built from two directions into the mountain West of the united states, meeting in Utah. if we can meet at the same place without too much effort, that would be significant in and of itself.

If the Land Shark here has not fully and honestly disclosed his methods, then pray tell what the methods in fact were, or had to be. The criminal mind no less than any other, wishes praise for perceived genius. Perhaps this criminal mind has deceived even his Confessors here.

I'm trying not get confused by the layers of double bluff here....

But if you were the guy that rigged the election: then you didn't simply distribute fraud randomly in both directions (erring a little in Bush's favour) or I'd have caught you.

So.... what you might have done is to avoid precincts where Bush was doing really well or really badly. What you did instead was to randomly pick precincts where Bush was doing OK, but not too brilliantly, and randomly select a lower limit to the proportion of votes to which you would let him fall. Not a consistent limit - that would show up - it would have to have a fairly natural distribution.

I'm going to have to model this. I'm still not convinced this is how you did it. I don't think you bothered with faking Kerry fraud, I don't think that would cover your traces at all, it would make it more obvious, and help Kerry at the same time.

The other possibility is that you planned all this in 2000, to give the Fraud Squad a fake baseline. Trouble is, you did it without shifting the exit polls much in 2000....

Boy, you were clever. But I'm not giving up yet....

...of yer strategy. He has a problem with the tabulator grabs & nefarious code which leave too many fingerprints for FBI boy scouts to find.

Cool post!!! Thx.

a step away from source code of one sort er another.

Please see Section 4 of the latest version of the RFP which clearly states that:

4.1 Precinct totals must ALWAYS match central tabulator totals.

4.2 A mass media diversion prior to the election must be created by an astro-turf activist group showing how easily votes can be switched AFTER the election at the central tabulators.

4.2.1 This will serve to divert attention from the impending precinct level discrepancies which will be engineered en masse BEFORE the election, or during early voting periods when the likely results will become apparent and subject to change without notice.

:thumbsup:

You suggest, it seems to me, that the fraud was committed so that recounts wouldn't be able to identify the fraud.

I don't believe that that was what happened in Ohio. There were only about 6 counties that used electronic voting. All the other ones have paper trails and therefore could have been recounted. Furthermore, I believe that there are still lawsuits going on to force those recounts, and other work as well is going on to force those recounts. And anyhow, if a recount wouldn't have identified the fraud, why did election officials go to such extremes (including two Cuyahoga County officials who have been indicted) to prevent a recount?

It seems to me that a recount of Ohio, even of selected counties such as Cuyahoga, provide our best chance at this time of proving a Kerry victory. I hope we don't lose sight of that fact.

...at this point I bet the result will be that ** won. Not that he really did, but the "trail" was tampered with from the get-go. The best hope of the lawsuits going forward is to prove that the Recount was not done lawfully, and to set precedent for future cases and elections.

In the first Recount last December, in numerous counties I saw with my own eyes, "randomly chosen" ballot boxes that were laid out on the recount table and waiting for us when we arrived to observe.

In at least one of those counties, I watched as the numbered seals were cut on those boxes (seals which were supposed to have been in place since Election Night), and inside were all the ballots sorted perfectly by who was voted on for President.

There was not even one hanging chad in the 250 ballots I personally observed, in fact the holes appeared to have been punched with laser beam razor sharp precision. Either Ohio has the best voter education in the history of the earth on how to punch votomatic cards, or 'them thar cards wuz tampered with'.

I'll bet the latter.

As Andy often sang, "Paper trails to you... so we can steal your vote..."

Proving in a recount that Kerry won, versus proving that the ballots in every single county were tampered with are two very similar things. I still think the recount needs to go forward.

That is very important information you provide here. I hope you get to testify in court about it.

I think the lawsuits definitely need to go forward, and I think the "paper trail" should be recounted at some point.

Just be aware that it is quite probable that if/when we do recount, based on what paper is still there now the results will show ** won. There were many reports of tampering, like what I saw or worse.

Hurrah, to definitively prove or disprove something: watch out.

Palast was saying early on spoilage was the key to Ohio. Add ten hour lines.

solid evidence of tampering.

If tampering could not be ruled out but could not be proven either, that would be a wash I guess.

rule out recount-detectable errors, but only if one believes the chain of custody was secure at all times and there was no ballot box stuffing

Perhaps the observers for the Greens-Libertarian December recount were not able to inspect the ballots sufficiently. But under the Sunshine laws, the observer can inspect each ballot and check the ballot number. The ballot numbers should correspond to those in the poll book. All over and under votes would need to be accounted for, also - and the votes counted wherever voter intent was at all clear.

In the Dec. recounts, over and under votes were NOT counted in any county I know of. Those neatly sorted ballots you saw had obviously had the under and over votes removed ahead of time.

HERE'S THE GAME/LEGAL ATTACK: IT WAS EXPRESSLY LITIGATED IN THE ELECTION CONTEST IN WASHINGTON STATE:

the Rs argue that the statute authorizes a Re-COUNT, not a re-CANVASS. Canvassing is the process of determining whether a ballot's intent is and whether it is entitled to be counted under registration laws. Counting is counting.

So, any ballots improperly, illlegally, and arbitrarily rejected as spoiled the first time around ARE NOT ENTITLED TO BE RE=COUNTED because they were never COUNTED in the first place!

Beauty!

Elegance!

No safety net for voters! Foolproof Crime! Sanctioned by the courts and based on the tiniest details of statutes thought out long in advance for suppression purposes, then consciously argued expressly to suppress votes from being counted and to prevent mistakes from being caught.

The SECOND appeal to the WA Supreme Court allowed more ballots to be counted, but only because the canvassers WANTED to, in that they had neglected to count them the first time. "administrative oversight" is sufficient to get them considered, but only if this reason is initiated by the county government itself.

Sounds like this is exactly what happened in Ohio. Wouldn't this really constitute "tampering" of a great degree, but *not* detectable under ReCOUNTS?

It doesn't sound as though Washington state has Sunshine laws.

Florida does have Sunshine laws, but because of their intensive use of paper-less DREs, there are no ballots to examine.

but they were either not argued or not argued strongly enough by the Dems. But you can look at it through sunshine, the question is whether or not it will be counted. And there were/are specific statutes covering that, which are likely to prevail over general sunshine laws.

that the 3% manual recount was conducted properly. Apparently the BoEs were not violating the law when they refused to examine the under/over votes. But the Ohio recount occurred AFTER the count was certified, so it wouldn't have affected the count anyway.

Recounts under the Sunshine laws can't change the official count either, of course, but they would tell us much more about what happened in Ohio.

if they had ever been allowed properly. This is because of the relative lack of DREs in Ohio (but not total lack) and the Blackwell decision to stay with the game he knew for 2004. As I said, the technology combined with the administrative procedures will dictate the types of attacks that are most likely.

However, for purposes of my explanation above, it is clearly underappreciated by folks that many forms of election rigging are not recount detectable or exit poll detectable and won't be found in source code TODAY, or perhaps at any time. Therefore, while information should be obtained, caution should be exercised in setting up any information as the holy grail of fraud research, because it's not.

A good election system creates evidence and witnesses on attack, provides a low payoff per crime (such as a single measly vote), and has inherent security precautions such as physical size whenever bulk fraud is attempted, and has time-limitations such that it can only be accomplished on election day (true of paper ballots that have detachable serial numbers issued in finite numbers to recorded polling places)

A bad election system, but one the Land Shark loves for fraud purposes, leaves little evidence or witnesses (none if you do it right), has a high payoff per attack (perhaps even the entire election) and can be accomplished at leisure even before election day, and has ballots that can be changed or destroyed noiselessly and soundlessly. No smell when burning. (Electronic voting) If the chosen method is self deleting code that activates only on election day, the code can be added at leisure to the machine before election day, greatly reducing the hectic stresses that fraudsters otherwise have on Election day.

..."self deleting code" as it can't delete backups or every place the software ever existed on the hard disk. Why leave fingerprints when my inner-fraudster can imagine & implement plenty of pre-planned phukupz? My inner-fraudster shure hatez lieberalz, but luvz freedom too much to leave any trace.

and put some executable code on it. That's all you need. Then it can be deleted. So, anyone who would have access to a single machine any time on election day or prior could get a good payload code right into the tabulator. But that's tab fraud.

Who says there are finger prints? Do you think Visual Basic, properly handled, will leave fingerprints?

BTW, whose gonna know? Homeland security specifically announced the Diebold back door as a cyber-security alert PRIOR to the election. Don't you think other criminals knew about that? Apparently Homeland Security thought enough criminals knew or could know that a published warning was thought necessary (or some other reason) You might find a little evidence. nobody's looking. Most particularly elections officials don't want any embarrassment and have signed contracts pledging to help defend corporation trade secrecy through "any lawful means". Does this sound like whistleblowing authorization to you? If a trade secret is anything that might impact one's competitive position, wouldn't you as corporate vendor take the position that the footprints of a hacker were somehow trade secrets?

...entails far more than a simple file deletion to eradicate all fingerprintz.

Even with free rein of the system, acting with impunity requires controlling the evidence fer a long long time. Not sure my inner-fraudster trusts all the other inner-fraudsters quite that much, but ya never no. If they paid me enuf, I'd prolly be retyred & sailing the whirled before the boy scouts could ever catch on.

Do you know of a single test for rigged code that is performed?

An elections official made the statement "our machines don't have any communicatoins capability". when pressed, the official admitted (1) he's never looked (2) they don't test and (3) nobody else does either. There was no rational basis for the comment. (trusting vendor statements I guess)

But the official "knows" and insists this is true. And these categorical denials get picked up and trusted by the media all the bloody time.

Not only is the Op good, so are the responses.

1. The method MUST provide a NET BENEFIT to * in terms of votes, at least in the targeted Electoral states, and is EXTREMELY LIKELY to provide a net benefit nationwide to insure popular vote victory.

So far, so Rove.

2. Expectations are all formed by the immediately prior presidential election, which the 2004 election is shaping up as a replay of. I MUST stay within spitting distance of Bush2000's percentage performance, or have plausible cover for exceeding it.

Presidential rematches are typically landslides, not a rehash of the previous outcome; the criminal mastermind would aim for Nixon '72 or Reagan '84, not another go under the microscope (re: Equalize #1).

But then he has to do it without getting caught, if his client has Bush's approval ratings. Getting caught is the devil, losing the election is the deep blue sea. Life in the free trades.

3. If my method involves directly distorting/changing the actual reported or recorded election results as opposed to the suppression of voters prior to actually voting, the method MUST use actual votes as an input or else pollbooks showing the number of voters will be way too far away from actual counts of ballots.

Probably a good (evil) idea, but it puts a ceiling on non-undervote screwings.

4. If computerized, the fraud method would BE EXTREMELY LIKELY TO use a randomizer of some sort, in order to avoid detection by well known mathematical tests for randomness.

The random behavior of the voter is all that's required. Number generators are fairly costly in terms of covering tracks, and programmers can barely find their own bugs.

The best way to avoid detection is to make a DRE interface equally inscrutable to all voters (in a high-Kerry precinct, like Philly or Broward). Stacking the deck is the easiest way to cheat without sleight-of-hand (re: #2 spray): "The law, in its majestic equality, forbids the rich as well as the poor to sleep under bridges, to beg in the streets, and to steal bread."

5. My PREFERRED METHOD FOR ELECTION FRAUD WOULD BE VERY LIKELY to mix it up a little, providing some benefits for the opponent Kerry so long as the net benefit to the favored candidate *.

Scylla and Charybdis again: if Florida 2000 was decided by a few hundred votes, evil mastermind couldn't afford to give quarter.

6. To deal with exit poll problems, I would want my FRAUD TO MASQUERADE AS POLLING BIAS, (that age old bugaboo of all polls) thus totally cutting off any enterprising (and oh so PARTISAN) election activists at the knees with an army of very able and honest and completely noncomplicit statisticians and political scientists who will sincerely argue that the exit poll data is more consistent with polling bias than anything else.

This is where the mastermind would need a team of MIT grads with nothing better to do prank-wise (although it explains why the highest WPEs are within driving distance of Cambridge).

(1) "ROUGHLY EQUALIZE TO 2000, NET AT 2000% plus 0.4%" or something like 0.4%.

Why 0.4%, though? If Mitofsky gave a +-3% on the state polls, would that make <3% fraud noiseless from the overmind perspective?

(2) VOTE SWITCHING AND SPRAYING NONRANDOM PRECINCTS... thus scattering the margin provided by corrupted machines over the entire polling place MIMICKING A LEGITIMATE VICTORY because the whole polling place "trends" Republican.

To my knowledge, every report of DRE switching occurred in historically +Dem polls/precincts. Statistical spraying is way too abstruse for the likes of Rove, it's probably a stacked deck: the rigged DREs could switch all candidates' votes ("miscalibrate"), but if the machines themselves are concentrated in deep-blue counties the effect is the same. It's nearly impossible to keep "easter egg" code hidden, but unconscionably bad design is universal (fits in with the noise). The "Clint Curtis hack" is about tabulators, precisely because vote-switching software on the machine level would be too widely disseminated to keep secret.

(3) DEMOCRATIC VOTES ARE NOT RECORDED OR DELETED in high minority areas

This is the second best way, since minorities tend to be disenfranchised from the legal/political establishments that examine these disputes in the first place. It has the negative side effect of turning people 98-2% against your candidate, but that doesn't matter thanks to (#4):

(4) In recountable areas that will be close, TIME-HONORED VOTE SUPPRESSION TACTICS

worth repeating, in no particular order:

Mixing it up won't hide it from the new correlation. Fraud in the blueshift direction will correspond with swing away from Bush just as Fraud in the redshift direction will correspond with swing towards. Both will introduce the same correlation, only the intercept will be different (to Bush's disadvantage) so there's no point that I can see.

Detecting randomness is easy. It would be harder to detect non-random algorithms that were designed to interact with votes cast and make an "intelligent" decision as to whether to switch/delete/add votes.

Fraud masquerading as polling bias is difficult for one simple reasons:

Polling bias results only in redshift. Not in more votes for Bush.
Fraud results in redshift AND more votes for Bush.

There is no real way round that. Well, there may be lots of ways, but they will have to be pretty clever. Because there are various fingerprints that will tell you the difference unless you take care to wear gloves:

1. Polling bias will cause affected precincts to have increased redshift - let's call that direction "up".

Fraud will cause affected precincts to move not only "up" but "rightwards" (literally and metaphorically) along the spectrum of Bush's vote share. Therefore a swarm of precincts from the centre of teh plot will tend to move diagonally relative to the rest, and introduce a positive correlation between redshift and Bush's vote share. There does not appear to be such a correlation (pace Ron Baiman).

2. Fraud will cause affected precincts to have greater swing relative to 2000. This will introduce a correlation between red shift and swing. There does not appear to be such a correlation.

3. Fraud will cause affected states to have greater swing relative to pre-election polls than unaffected states. State-level red shift should therefore correlate positively with deviation from pre-election polls. OTOH has, I think, tested this, and found that if anything the correlation is negative.

All these follow from the necessary premise that fraud has to result in benefit to Bush. If fraud was the cause of Bush's win, then the fraud must have managed to intelligently masked its own fingerprints. I don't think we've quite figured yet how that could have been done (but I'm working on it....)

Granted the perp would need some working knowledge of statistics to pull off the great Exit Poll Heist of Vermont, but the criminal we're empathizing with doesn't share our hindsight.


It would be hard to detect statistically, but easy to detect forensically. Tabulator fraud has the opposite problem.

Why are you admitting this here, instead of to a prosecutor or to someone in the press? If you did all this stuff, and are willing to talk about it, why not come out?

Of course the delay for conversation is their undoing and Batman gets away, but that last part is just TV.

The press isn't listening, BTW.

Prosecutors: You can count the number of prosecutors pursuing white collar crime these days on your fingers. One reason DeLay probably believes in his gut he really is being victim of a witchhunt is that nobody is really going after white collar crime, even after Enron and the whole bit.

"In the trench warfare where suppression is necessary or the balloting systems less penetrable or diverse such as Ohio, the discrepancies created are not detectable by a recount, nor are they detectable by exit polls because voters are suppressed from seeing exit pollsters as well." -LS

if the voter is leaving in disgust without having voted, NO EXIT POLL DATA!!!!!

:rofl:

Moreover, check out my post above on recounts vs. recanvasses.(post 56) At least under the WA statutory structure, if votes aren't counted the first time, they don't *ever* count 'em unless they feel like it AND through administrative oversight they were not counted OR canvassed in the first count.

What's really funny is the naive americans who have such quaint faith that "recounts" will reveal the truth!

:rofl:

I suspect voter suppression was the fraud of choice in Ohio - and was responsible for the Electoral College selection of *.

If someone gets busted, they want to deny. If denial no longer works, then minimize. If minimization no longer works, then argue it doesn't change the result. If changing the result appears possible go nuclear with partisan hack sore loserman and launch all missiles to argue Kerry fraud to neutralize any bad news then move in for the kill with legal "importance of finality" "stability of government" and others enshrined in statutes as state public policy while the public wasn't really thinking it through.

But then, even if suppression is proven intentional, there should be enough other strategies to always fall back on "doesn't change the result, we won" line of attack.

Most of these occurred primarily in Democratic strongholds (urban precincts):

<> Long lines caused by too few machines and machine breakdowns

<> Voter intimidation

<> Failure to issue absentee ballots

<> Failure to issue provisional ballots in accordance with the law

<> Giving voters incorrect information in order to prevent them from voting

<> (Florida) Felon lists

<> Destruction of voter-registration forms

<> Comprehensive purges of occasional voters

<> Stricter voter-ID requirements


(Many of these are from a list by Tfc)

useful things, but not suppression, and suppression is a major part of the Ohio story.

I think it was Keats who had the notion of "negative capability" in the sense that the mark of intelligence is the ability to hold two opposed ideas in mind at the same time.

the negative capability concept for ohio is this: do the recounts for what they may reveal, but don't build them up as an answer, it's only some data. Like passing a breathalyzer doesn't mean the person isn't on drugs, and each drug test screens for one drug or one class of drugs

I've been doing a lot of that lately ;)

Coming to the conclusion that voter suppression is the best bet for stealing elections, but small quantities of other stuff spread about a bit would help it on its way, not too much of any one thing. Just a hint of a fingerprint here and there, nothing to stand out. Larger quantities of any one thing, I'm not so sure yet....

it means that even if caught, any one method "not enough to change the result". Media goes back to sleep..... Zzzzzzzzz.

But, it could also indicate many different independent efforts, none of them necessarily tied together but effectively acting in concert because of the fact that all partisans of all parties know Who What When Where and Why of what needs to be done, as does the entire public. The only question to be solved is How.

"...and suppression is a major part of the Ohio story." -LS

Voter suppression is the perfect crime.

Undetectable by exit polls or by recounts.

And if you can do it by allocating voting machines on the basis of past turnout, you have complete plausible deniability ("we tried to allocate the machines to precincts likely to need them most - we even threw in some extra ones in traditionally low-turnout precincts in case there was a greater than expected increase....)

I do think there is a slam dunk Civil Rights case in Franklin County (but that's based on my knowledge of stats not of law!)

on machine allocation in Franklin County,

but here's a plug for those who haven't:

http://uscountvotes.org/ucvAnalysis/OH/FranklinCountyReport_v2.pdf

Also - looked at your Twin Peaks. There are arguably three if you count the PB peak as yet another population (on edit: its mode is further to the right of the plot than the rightmost mode in the Twin Peaks plot, I think).

I agree with OTOH that the Ns are small for anything definitive but the data are certainly consistent with three "populations" - the question then is, are they populations of people or populations of counting methods?

a sample size equal to 100% of the population being tested?

If so, at least in this case, N size is not controlled by the researchers.

You are right that the shape is what it is. But if you want to compare your distribution with that of a normal distribution then the N matters. The question is: is your Twin Peaks distribution significantly more platykurtic than a normal distribution? (perfectly good statistical test), ie it different from the shape we'd expect under the null hypothesis, which is that all votes have an equal chance of being counted, regardless of method (terrible to think that chance has anything to do with it). And I'm not sure that with your N you have the power to do that.

OTOH will tell me if I'm talking rot.

There is an entire metaphysics devoted to statistical inference based on populations. (I am reminded of more innocent times, when I defended USCV's statement that response rates were higher in Republican precincts against some fellow named Rick Brady who insisted that if the difference wasn't statistically significant, it wasn't a fact. He still thinks he is right, and he could be, depending on how one interprets the statement. I argued that it was defensible as a merely descriptive statement. And, after all, one ought to be able to make descriptive statements about response rates, which by default are not inferential: response rates are ordinarily not reported with margins of error! But USCV was at least hinting at a possible inference.)

Anyone can say that the distribution is twin-peaked, no problem. But if we're gonna interpret that fact, then I don't see how we can escape statistical inference.

Platykurtic is a perfectly good test, but maybe not the best one. I would not say that there are three different populations here (I don't think I'm actually disagreeing with your post #74 -- EHEM! -- just shifting the rhetorical emphasis). If there are two distinct (sub)populations in the second plot, they constitute the population in the first plot. So if some process is affecting the electronic vote in some precincts and not in others, that deviation should (or might) be amenable to precinct-level analysis. I'm not sure what the test should be, but we probably can come up with a reasonable one that has more power (and offers more insight) than just comparing distribution moments, which throws out a lot of information.

(EDIT: malapropism)

floaty hearts.. I feel so.. so.. Betraayyed....( need a southern belle fainting smilie here..)

Where do we go from here? What do we do to out smart them in 06? How can we secure the machines? Seems from what you are saying (and I know you are right) even if we surround all the vote counters and central tabulators they still can cheat. How do we secure the machines with the laws that are in place right now (because we know the Carter commission)or the GAO ideas are not going to be in place for 06. Carters ideas from the 2000 election commission where not in place for 2004. In Carters words not mine

By Jimmy Carter
Monday, September 27, 2004; Page A19

After the debacle in Florida four years ago, former president Gerald Ford and I were asked to lead a blue-ribbon commission to recommend changes in the American electoral process. After months of concerted effort by a dedicated and bipartisan group of experts, we presented unanimous recommendations to the president and Congress. The government responded with the Help America Vote Act of October 2002. Unfortunately, however, many of the act's key provisions have not been implemented because of inadequate funding or political disputes.

The disturbing fact is that a repetition of the problems of 2000 now seems likely, even as many other nations are conducting elections that are internationally certified to be transparent, honest and fair.

http://www.washingtonpost.com/wp-dyn/articles/A52800-2004Sep26.html

Someone needs to do a cost analysis of what HAVA compliance would actually cost. Five million for our county doesn't even count half the vote, with 45 minute lines during rush periods. Our county is looking to cut its budget, not spend more. HAVA grants are there, but an illusion in the sense that they don't come close to funding the partially funded (non)mandate here.

OK, so
(1) establish the insane cost of this if truly implemented

(2) note the county clerks have just written to congress begging for 1/1/06 mercy (victims of their own or vendor-driven misinterpretation of HAVA requirements, which do not require technology change wholesale, they require disabled *access*)

(3) push for supplemental paper ballots in all polling locations that have DREs both to save taxpayer $$ as well AS AN ANSWER TO LONG LINES CREATED BY DRE BOTTLENECKS since there's not enough $$ to pay for all machines to cover an entire county anyway. THE PRINCIPLE HERE IS REDUNDANCY CREATES SECURITY. Assign a certain **random** number of voters (up to 50%) to paper at the polling places themselves without letting the voters choose paper or plastic, and then we have a study to check the accuracy of the voting machines (i.e. side by side paper and electronic, with NO VOTER CONTROL or choice as to which one they get) This forces an effective audit whether they like it or not, but does so in the name of saving taxpayer $$ and eliminating voting lines.

(4) If paper results differ from electronic, DO OVER.

On Edit: Some advantages are: The electronic results will do their part to speed up results. It gives election officials fig leafs for their ill-advised purchasing decisions because they will still be used, while also (subject to tweaking here) providing the data (with additional details) that will allow detailed statistical studies. Per other comments in this thread, it may also require that small precincts should not exist (besides the fact that really small precincts can inadvertently disclose voter identity through processes of elimination, when looking at published precinct results).

but I still believe it:

http://www.dailykos.com/storyonly/2004/12/19/9056/8130

then we tell THEM how the elections going to be run!

necessarily mean survival of the most evolved.

My copyrighted darwin fish bumper sticker features the darwin fish being swallowed by the Jesus fish who is in turn being engulfed by a shark labeled "Religious Right", taking a both/and view on the whole debate and still touching, I think, the right chord.

In the end, though, each animal has its evolutionary purpose and ecological niche.

These muther fucken vote stealing people who think that they can just roll over our asses while they are stealing our fuckin elections have got another thing coming. It's time to stop playing NICE with these Bastards and deal with these Cock suckers one on one right NOW, not after 06, its time NOW to kick there fuckin asses.

Now that we know what these bastards are fuckin doing, there bulshit has got to be stopped NOW. Lets FIND A WAY TO SHUT THEM DOWN, NOW!

Insist on a 5 minute voting service guarantee for all voters in all elections, assuming the turnout is equal to Nov 2004. then assume all voters vote electronically.

if they don't like assumption #2 (all voting electronically) then it's patently obvious that the only way we can vote electronically is with a SUBSIDY from absentee voting. In our county (68% paper absentee) the only way $5M can support even 32% of the vote is because the majority have voted with their feet to paper! In fact, the ONLY WAY WE CAN HAVE E-VOTING IS ESSENTIALLY BECAUSE WE ***DO NOT*** have ELECTRONIC VOTING!!!! (for the vast majority)

Answer: not enough money exists is almost all jurisdictions to do this five minute service guarantee, even if we lower the percentage voting electronically to 70% of all voters, for example.

So, the convenience angle on which DREs are sold is an outright FRAUD. It's a formula to create (this one's for you kster) FUCKING LINES, is what it is.

Then find out if the contracts proposed or actual in your county contain warranties. They are probably attempting to disclaim all "implied warranties of merchantability or fitness for a particular purpose". These implied warranties are in every sale transaction as a matter of law because no reasonable person purchasing from a merchant would want to pay a good price for something that isn't warranted to work "for its ordinary purposes" (merchantability) or for the particular purpose intended by the customer if the merchant has reason to know that purpose (fitness for particular purpose warranty).....

SO........ what this means is that the voting company in question is DENYING THAT THE THINGS WILL WORK FOR ANYTHING AT ALL, SPECIFIC OR GENERAL.

Is that what the voters want???

Long lines and no promises????

Is this what democracy deserves???

I've said many times before that when people wake up to the fact that not only does the emperor have no clothes, he has no wardrobe and lives in a glass house, it's all over.

As soon as people "own" and feel democracy even a bit like their own personal property, they will cast out the moneychangers from the temple of democracy pretty damn fast.

All that needs to happen is people need to remember and feel what democracy is all about, and then they will be PISSED ROYAL.

Facts are not needed so much as clarity, and passion.

Is this DU or DKos? Can I use bad words too?

DREs are a terrible idea if for no other reason than their cost.

Something that needs to be cheap and simple should never be made expensive and complicated, otherwise it will end up being rationed.
And when something is rationed, the powerful get more of it, and the poor get less.

And when what the poor get less of is access to the mechanism for selecting their government, you don't have a democracy.

Paper, pencil and people are all you need to conduct an election.

Why Pay More?

(Unless you WANT to disenfranchise the poor, of course...)

Sometimes things need to be just a bit more intense. Sometimes Roget's thesaurus falls just a bit short. Damn right about that.

instead of Alberto?

I'm really quite fond of this thread.

I think it's one of the most productive "fraud" threads I've seen.

Bravo, Land Shark!

and so I'm taking it out again tonight.

Febble was cute from the beginning but before too long OT-OH (translated now as Overtime-Alcohol) it seems would be just fine too. Peace. Love. And Understanding. Bartender, give me a double!

The shark bites!


:applause:

Thank you.

The hint of blood in the water.